Biomedical Engineering Reference
In-Depth Information
left ventricular myocardium. The scan is usually performed in two parts:
a. A scan at resting work rate and therefore basal myocardial perfu-
sion, and
b. A scan at peak stress where the radiotracer is injected as the patient
is undergoing a paradigm designed to increase myocardial perfusion,
either by physical exertion or pharmacological stress on the heart.
A common clinical question asked of this scan is whether areas of decreased
blood flow at stress are normally perfused at rest: this ''reversibility'' indi-
cates compromised blood flow in the arteries supplying the myocardium, but
normal underlying myocardial tissue. This may be improved by surgical
intervention. The fact that the rest scan is normal indicates that the underly-
ing myocardium is still functionally viable. In this example, therefore, the
study at stress and the study at rest need to be realigned to detect differences.
Visual inspection of the aligned images is often sufficient to report the scan.
However, a valuable tool to assess the change quantitatively is to compare
both the stress and rest studies with normal databases from subjects who
have undergone the same scan but with proven low likelihood of any myo-
cardial perfusion deficit. As the shape and size of different individuals' hearts
will vary, a spatial normalization step is required to “transform” each heart
to a standardized representation.
To illustrate this, a pair of stress
rest scans of a subject are shown in Figure 11.11.
The data are from a SPECT study using -tetrofosmin. The top rows
show the stress and rest scans coregistered using an automated routine.
99m
[
Tc
]
51
Next, myocardial perfusion throughout the left ventricle is determined and
plotted as a function of angle around the circular sections of the myocar-
dium seen in the scan. In an idealized scan, a profile generated over 360
around the myocardium in the short axis views would produce a constant
count. Due to normal variations in myocardial thickness and perfusion the
profile is not necessarily constant. This process is repeated for all of the
slices through the short axis projections and therefore the entire left ventric-
ular myocardium is mapped to a series of profiles of perfusion. A common
method for displaying these profiles in a compressed manner is the so
called “polar plot” display.
52
Each profile is mapped from the base of the
heart to the apex onto a series of concentric rings as if the three-dimensional
ellipsoid of the left ventricle has been “squashed.” Once in this format, the
data are now represented in a standard, spatially normalized space and can
be compared against other individuals' scans. These can be seen in the lower
pair of images in Figure 11.11. Some larger clinical centers have generated nor-
mal databases in this manner which consist of a series of subjects' count
profiles.
53
Individual subjects can be compared statistically against the normal
population profiles to objectively assess changes which may be outside the nor-
mal range. This is one of the first examples in nuclear medicine of the use of spatial
transformations for cohort studies. By virtue of the simplification of a three-
dimensional ellipsoidal distribution into a single image, the polar plot has
